With the continuous improvement of science and technology related to food olfactory and taste senses and emotional cognition， more and more instrumental analysis methods and experimental equipment are used in the research of the above fields. The diversification and comprehensiveness of detection methods and the improvement of detection accuracy are accompanied by the expansion of data scale related to flavor perception. How to obtain key information from the analysis results of food flavor instruments and the large amount of data collected in the research of consumers' emotional cognition and behavior， and establish the correlation between the data， has been paid more and more attention by researchers. Data mining and modeling technology in the field of food is to use a large amount of data obtained in the process of food production and circulation to monitor the physical and chemical changes in each link of the food industry chain in real time and accurately， and to predict the impact of these changes on consumers' sensory characteristics and emotional cognition. In the field of food smell and taste perception， data mining and modeling techniques can provide unprecedented insight and analytical power to food researchers and consumers. Based on the supervised and unsupervised data mining and modeling methods and deep learning methods of classical machine learning methods， this paper analyzes the latest application progress in the research of food sensory attributes and emotional cognition， and looks forward to the application prospect of data mining and modeling technology in the field of food olfactory and taste perception. Help the scientific and technological progress and industrial upgrading of the food industry.